An AI photo manager that reads your train tickets and ranks your selfies

What it does

TrailSnap is a self-hosted photo management system that indexes images by location, face, scene, and time. It extracts metadata to build footprint maps, timelines, and conditional albums, and uses OCR to pull trip details out of physical train tickets and itineraries. The stack ships as Docker containers with a web frontend and Python backend.

The interesting bit

The author claims over 90% of the code was AI-generated using Trae and Claude Code, which explains both the ambitious scope and the slightly scattered roadmap. The project also targets distinctly Chinese travel habits, like tracking visits to 5A-rated scenic spots and parsing domestic railway tickets.

Key highlights

• Supports face recognition, scene classification (pets, food, night shots), and smart search by content or person.
• Generates an annual travel report with photo walls, city counts, and route mileage.
• Offers an “On this day in previous years” view that auto-plays historically memorable photos.
• Can index external folders and handles Live Photos from iPhone, vivo, OPPO, and Xiaomi.
• Exposes agent and skill interfaces for OpenClaw and Claude Code integration.

Caveats

• Several headline AI features—auto-edited VLOGs, AI photo retouching, and comprehensive travel diaries—are still marked as pending in the roadmap, despite some agent tooling already being present.
• The README contains conflicting signals about which AI capabilities are currently live versus future work.
• Some itinerary features, like concert or movie ticket management, remain under development.

Verdict

Worth a look if you want a self-hosted Google Photos alternative with a travel bent and don’t mind sorting through a few work-in-progress features. Skip it if you need a mature, fully realized AI media pipeline today.